arXiv subjects

Bo Sun

Publications and source records attributed to Bo Sun.

At least 109 records · Page 6Linked to original sources

arXiv2020

H3DNet: 3D Object Detection Using Hybrid Geometric Primitives

We introduce H3DNet, which takes a colorless 3D point cloud as input and outputs a collection of oriented object bounding boxes (or BB) and their semantic labels. The critical idea of H3DNet is to predict a hybrid set of geometric primitives, i.e., BB centers, BB face centers, and BB edge centers. We show how to convert the predicted geometric primitives into object proposals by defining a distance function between an object and the geometric primitives. This distance function enables continuous optimization of object proposals, and its local minimums provide high-fidelity object proposals. H3DNet then utilizes a matching and refinement module to classify object proposals into detected objects and fine-tune the geometric parameters of the detected objects. The hybrid set of geometric primitives not only provides more accurate signals for object detection than using a single type of geometric primitives, but it also provides an overcomplete set of constraints on the resulting 3D layout. Therefore, H3DNet can tolerate outliers in predicted geometric primitives. Our model achieves state-of-the-art 3D detection results on two large datasets with real 3D scans, ScanNet and SUN RGB-D.

cs.CV↗

arXiv2020

Delay-Aware Scheduling over mmWave/Sub-6 Dual Interfaces: A Reinforcement Learning Approach

We consider a transmitter with mmWave/sub6 dual interfaces. Due to the intermittency of mmWave channel, the transmitter must schedule packets wisely across the interfaces to minimize the average delay by observing the system state. We usethe well-known dynamic programming methods and Q-learning to find the optimal scheduling policy and investigate the influenceof observing CSI on the optimal policy under different levels of knowledge of the environment. We find that only when the channel state transition model is not available, the instantaneousCSI can help in reducing system delay

cs.NI↗

arXiv2020

Modeling Multi-Cellular Dynamics Regulated by ECM-Mediated Mechanical Communication via Active Particles with Polarized Effective Attraction

Collective cell migration is crucial to many physiological and pathological processes. Recent experimental studies have indicated that the active traction forces generated by migrating cells in fibrous extracellular matrix (ECM) can mechanically remodel the ECM, enabling long-range propagation of cellular forces and leading to correlated migration dynamics regulated by the mechanical communication among the cells. Motivated by these experimental discoveries, we develop an active-particle model with polarized effective attractions (APPA) for modeling emergent multi-cellular migration dynamics regulated by ECM-mediated mechanical communications. Active particles with polarized pairwise attractions exhibit enhanced aggregation behaviors compared to classic active Brownian particles, especially at lower particle densities and larger rotational diffusivities. Importantly, in contrast to the classic ABP system, the high-density phase of APPA system exhibits strong dynamic correlation, which is characterized by the slowly decaying velocity correlation functions with a correlation length comparable to the linear size of high-density phase domain (i.e., cluster of the particles). The strongly correlated multi-cellular dynamics predicted by the APPA model are subsequently verified in {\it in vitro} experiments using MCF-10A cells. Our studies also indicate the importance of incorporating ECM-mediated mechanical coupling among the migrating cells for appropriately modeling emergent multi-cellular dynamics in complex micro-environments.

cond-mat.soft↗

arXiv2020

First-Principles Study of Hydrogen Behaviors in $α$-Pu$_{2}$O$_{3}$

The in-depth understanding of hydrogen permeation through plutonium-oxide overlayers is the prerequisite to evaluate the complex hydriding induction period of Pu. In this work, the incorporation, diffusion and dissolution of hydrogen in $α$-Pu$_{2}$O$_{3}$ are investigated by the first-principles calculations and $\textit{ab initio}$ thermodynamic method based on DFT+U and DFT-D3 schemes. Our study reveals that the hydrogen incorporation is endothermic and the separated H atoms prefer to recombine as H$_{2}$ molecules rather than reacting with $α$-Pu$_{2}$O$_{3}$. The H and H$_{2}$ diffusion are both feasible, generally, H will recombine first as H$_{2}$ and then migrate. Both pressure P$_{H2}$ and temperature can promote the hydrogen dissolution in $α$-Pu$_{2}$O$_{3}$. The single H$_{2}$ molecule incorporation and (H+H$_{2}$) mixed dissolution will successively appear when increasing P$_{H2}$. Compared to PuO$_{2}$, this work indicates that Pu sesquioxide is hardly reduced by hydrogen, but the porous $α$-Pu$_{2}$O$_{3}$ facilitates hydrogen transport in Pu oxide layers. We presents the microscopic picture of hydrogen behaviors in the defect-free $α$-Pu$_{2}$O$_{3}$, which could shed some light on the study of the hydriding induction period of Pu.

cond-mat.mtrl-sci↗

arXiv2019

Adversarial Defense by Stratified Convolutional Sparse Coding

We propose an adversarial defense method that achieves state-of-the-art performance among attack-agnostic adversarial defense methods while also maintaining robustness to input resolution, scale of adversarial perturbation, and scale of dataset size. Based on convolutional sparse coding, we construct a stratified low-dimensional quasi-natural image space that faithfully approximates the natural image space while also removing adversarial perturbations. We introduce a novel Sparse Transformation Layer (STL) in between the input image and the first layer of the neural network to efficiently project images into our quasi-natural image space. Our experiments show state-of-the-art performance of our method compared to other attack-agnostic adversarial defense methods in various adversarial settings.

cs.CV↗

arXiv2019

Modeling cell migration regulated by cell-ECM micromechanical coupling

Cell migration in fibreous extracellular matrix (ECM) is crucial to many physiological and pathological processes such as tissue regeneration, immune response and cancer progression. During migration, individual cells can generate active pulling forces via actin filament contraction, which are transmitted to the ECM fibers through focal adhesion complexes, remodel the ECM, and eventually propagate to and can be sensed by other cells in the system. The microstructure and physical properties of the ECM can also significantly influence cell migration, e.g., via durotaxis and contact guidance. Here, we develop a computational model for cell migration regulated by cell-ECM micro-mechanical coupling. Our model explicitly takes into account a variety of cellular level processes including focal adhesion formation and disassembly, active traction force generation and cell locomotion due to actin filament contraction, transmission and propagation of tensile forces in the ECM, as well as the resulting ECM remodeling. We validate our model by accurately reproducing single-cell dynamics of MCF-10A breast cancer cells migrating on collagen gels and show that the durotaxis and contact guidance effects naturally arise as a consequence of the cell-ECM micro-mechanical interactions considered in the model. Moreover, our model predicts strongly correlated multi-cellular migration dynamics, which are resulted from the ECM-mediated mechanical coupling among the migrating cell and are subsequently verified in {\it in vitro} experiments using MCF-10A cells. Our computational model provides a robust tool to investigate emergent collective dynamics of multi-cellular systems in complex {\it in vivo} micro-environment and can be utilized to design {\it in vitro} micro-environments to guide collective behaviors and self-organization of cells.

q-bio.CB↗

arXiv2019

Oriented Point Sampling for Plane Detection in Unorganized Point Clouds

Plane detection in 3D point clouds is a crucial pre-processing step for applications such as point cloud segmentation, semantic mapping and SLAM. In contrast to many recent plane detection methods that are only applicable on organized point clouds, our work is targeted to unorganized point clouds that do not permit a 2D parametrization. We compare three methods for detecting planes in point clouds efficiently. One is a novel method proposed in this paper that generates plane hypotheses by sampling from a set of points with estimated normals. We named this method Oriented Point Sampling (OPS) to contrast with more conventional techniques that require the sampling of three unoriented points to generate plane hypotheses. We also implemented an efficient plane detection method based on local sampling of three unoriented points and compared it with OPS and the 3D-KHT algorithm, which is based on octrees, on the detection of planes on 10,000 point clouds from the SUN RGB-D dataset.

cs.CV↗

arXiv2019

Collective dynamics and self-organization of active particles on random networks

Collective cell migration in 3D extracellular matrix (ECM) is crucial to many physiological and pathological processes. Migrating cells can generate active pulling forces via actin filament contraction, which are transmitted to the ECM fibers and lead to a dynamically evolving force network in the system. Here, we elucidate the role of such force network in regulating collective cell behaviors using a minimal active-particle-on-network (APN) model, in which active particles can pull the fibers and hop between neighboring nodes of the network following local durotaxis. Our model reveals a dynamic transition as the particle number density approaches a critical value, from an "absorbing" state containing isolated stationary small particle clusters, to a "active" state containing a single large cluster undergone constant dynamic reorganization. This reorganization is dominated by a subset of highly dynamic "radical" particles in the cluster, whose number also exhibits a transition at the same critical density. The transition is underlaid by the percolation of "influence spheres" due to the particle pulling forces. Our results suggest a robust mechanism based on ECM-mediated mechanical coupling for collective cell behaviors in 3D ECM.

physics.bio-ph↗

arXiv2019

The morphodynamics of 3D migrating cancer cells

Cell shape is an important biomarker that is directly linked to cell function. However, cell morphodynamics, namely the temporal fluctuation of cell shape is much less understood. We study the morphodynamics of MDA-MB-231 cells in type I collagen extracellular matrix (ECM). We find ECM mechanics, as tuned by collagen concentration, controls the morphodynamics but not the static cell morphology. We employ machine learning to classify cell shape into five different morphological phenotypes corresponding to different migration modes. As a result, cell morphodynamics is mapped into temporal evolution of morphological phenotypes. We systematically characterize the phenotype evolutions including occurrence probability, dwell time, transition flux, and 3D migrational characteristics. We find that manipulating Rho-signaling enhances the morphodynamics and phenotype transitions. Using a tumor organoid model, we show that the distinct invasion potentials of each phenotype modulate the phenotype homeostasis. Overall invasion of a tumor organoid is facilitated by individual cells searching for and committing to phenotypes of higher invasive potential. In conclusion, we show that 3D migrating cancer cells exhibit rich morphodynamics that is regulated by ECM mechanics, Rho-signaling, and is closely related with cell motility. Our results pave the way to the systematic characterization and functional understanding of cell morphodynamics as a new biomarker for normal and malignant cells.

q-bio.CB↗

arXiv2019

Collective ECM remodeling organizes 3D collective cancer invasion

Tumor metastasis, traditionally considered as random spreading of individual cancer cells, has been shown to also involve coordinated collective invasion in the tissue space. Here we demonstrate experimentally and computationally that physical interactions between cells and extracellular matrix (ECM) support coordinated dissemination of cells from tumor organoids. We find that collective remodeling of the ECM fibrous structure by cell-generated forces produces reciprocal cues to bias cell motility. As a result, dissemination of cells from tumor organoids are controlled by the organoid geometry. Our results indicate that mechanics and geometry are closely coupled in multicellular processes during metastasis and morphogenesis. This study also suggests that migrating cells in 3D ECM represent a distinct class of active particle system where the collective dynamics is governed by the micromechanical remodeling of the environment rather than direct particle-particle interactions.

physics.bio-ph↗

arXiv2019

Stochastic Replica Voting Machine Prediction of Stable Cubic and Double Perovskite Materials and Binary Alloys

A machine learning approach that we term the `Stochastic Replica Voting Machine' (SRVM) algorithm is presented and applied to a binary and a 3-class classification problems in materials science. Here, we employ SRVM to predict candidate compounds capable of forming stable perovskites and double perovskites and further classify binary ($AB$) solids. The results of our binary and ternary classifications compared well to those obtained by SVM and neural network algorithms.

cond-mat.mtrl-sci↗

arXiv2018

MVPNet: Multi-View Point Regression Networks for 3D Object Reconstruction from A Single Image

In this paper, we address the problem of reconstructing an object's surface from a single image using generative networks. First, we represent a 3D surface with an aggregation of dense point clouds from multiple views. Each point cloud is embedded in a regular 2D grid aligned on an image plane of a viewpoint, making the point cloud convolution-favored and ordered so as to fit into deep network architectures. The point clouds can be easily triangulated by exploiting connectivities of the 2D grids to form mesh-based surfaces. Second, we propose an encoder-decoder network that generates such kind of multiple view-dependent point clouds from a single image by regressing their 3D coordinates and visibilities. We also introduce a novel geometric loss that is able to interpret discrepancy over 3D surfaces as opposed to 2D projective planes, resorting to the surface discretization on the constructed meshes. We demonstrate that the multi-view point regression network outperforms state-of-the-art methods with a significant improvement on challenging datasets.

cs.CV↗

arXiv2018

The Stochastic Replica Approach to Machine Learning: Stability and Parameter Optimization

We introduce a statistical physics inspired supervised machine learning algorithm for classification and regression problems. The method is based on the invariances or stability of predicted results when known data is represented as expansions in terms of various stochastic functions. The algorithm predicts the classification/regression values of new data by combining (via voting) the outputs of these numerous linear expansions in randomly chosen functions. The few parameters (typically only one parameter is used in all studied examples) that this model has may be automatically optimized. The algorithm has been tested on 10 diverse training data sets of various types and feature space dimensions. It has been shown to consistently exhibit high accuracy and readily allow for optimization of parameters, while simultaneously avoiding pitfalls of existing algorithms such as those associated with class imbalance. We very briefly speculate on whether spatial coordinates in physical theories may be viewed as emergent "features" that enable a robust machine learning type description of data with generic low order smooth functions.

stat.ML↗

arXiv2018

Stay On-Topic: Generating Context-specific Fake Restaurant Reviews

Automatically generated fake restaurant reviews are a threat to online review systems. Recent research has shown that users have difficulties in detecting machine-generated fake reviews hiding among real restaurant reviews. The method used in this work (char-LSTM ) has one drawback: it has difficulties staying in context, i.e. when it generates a review for specific target entity, the resulting review may contain phrases that are unrelated to the target, thus increasing its detectability. In this work, we present and evaluate a more sophisticated technique based on neural machine translation (NMT) with which we can generate reviews that stay on-topic. We test multiple variants of our technique using native English speakers on Amazon Mechanical Turk. We demonstrate that reviews generated by the best variant have almost optimal undetectability (class-averaged F-score 47%). We conduct a user study with skeptical users and show that our method evades detection more frequently compared to the state-of-the-art (average evasion 3.2/4 vs 1.5/4) with statistical significance, at level α = 1% (Section 4.3). We develop very effective detection tools and reach average F-score of 97% in classifying these. Although fake reviews are very effective in fooling people, effective automatic detection is still feasible.

cs.CR↗

arXiv2018

Dislocation-induced thermal transport anisotropy in single-crystal group-III nitride films

Dislocations, one-dimensional lattice imperfections, are common to technologically important materials such as III-V semiconductors, and adversely affect heat dissipation in e.g., nitride-based high-power electronic devices. For decades, conventional models based on nonlinear elasticity theory have predicted this thermal resistance is only appreciable when heat flux is perpendicular to the dislocations. However, this dislocation-induced anisotropic thermal transport has yet to be seen experimentally. In this study, we measure strong thermal transport anisotropy governed by highly oriented threading dislocation arrays along the cross-plane direction in micron-thick, single-crystal indium nitride (InN) films. We find that the cross-plane thermal conductivity is more than tenfold higher than the in-plane thermal conductivity at 80 K when the dislocation density is on the order of ~3x10^10 cm^-2. This large anisotropy is not predicted by the conventional models. With enhanced understanding of dislocation-phonon interactions, our results open new regimes for tailoring anisotropic thermal transport with line defects, and will facilitate novel methods for directed heat dissipation in thermal management of diverse device applications.

cond-mat.mtrl-sci↗

arXiv2017

On Equivalence of Known Families of APN Functions in Small Dimensions

In this extended abstract, we computationally check and list the CCZ-inequivalent APN functions from infinite families on $\mathbb{F}_2^n$ for n from 6 to 11. These functions are selected with simplest coefficients from CCZ-inequivalent classes. This work can simplify checking CCZ-equivalence between any APN function and infinite APN families.

cs.CR↗

arXiv2017

Asymptotic Performance Evaluation of Battery Swapping and Charging Station for Electric Vehicles

A battery swapping and charging station (BSCS) is an energy refueling station, where i) electric vehicles (EVs) with depleted batteries (DBs) can swap their DBs for fully-charged ones, and ii) the swapped DBs are then charged until they are fully-charged. Successful deployment of a BSCS system necessitates a careful planning of swapping- and charging-related infrastructures, and thus a comprehensive performance evaluation of the BSCS is becoming crucial. This paper studies such a performance evaluation problem with a novel mixed queueing network (MQN) model and validates this model with extensive numerical simulation. We adopt the EVs' blocking probability as our quality-of-service measure and focus on studying the impact of the key parameters of the BSCS (e.g., the numbers of parking spaces, swapping islands, chargers, and batteries) on the blocking probability. We prove a necessary and sufficient condition for showing the ergodicity of the MQN when the number of batteries approaches infinity, and further prove that the blocking probability has two different types of asymptotic behaviors. Meanwhile, for each type of asymptotic behavior, we analytically derive the asymptotic lower bound of the blocking probability.

cs.PF↗

arXiv2017

A Study on the Vulnerabilities of Mobile Apps associated with Software Modules

This paper reports a large-scale study that aims to understand how mobile application (app) vulnerabilities are associated with software libraries. We analyze both free and paid apps. Studying paid apps was quite meaningful because it helped us understand how differences in app development/maintenance affect the vulnerabilities associated with libraries. We analyzed 30k free and paid apps collected from the official Android marketplace. Our extensive analyses revealed that approximately 70%/50% of vulnerabilities of free/paid apps stem from software libraries, particularly from third-party libraries. Somewhat paradoxically, we found that more expensive/popular paid apps tend to have more vulnerabilities. This comes from the fact that more expensive/popular paid apps tend to have more functionality, i.e., more code and libraries, which increases the probability of vulnerabilities. Based on our findings, we provide suggestions to stakeholders of mobile app distribution ecosystems.

cs.CR↗